Appendix: Further details on the background correction models and their implementation

This appendix provides further specifics on the models described briefly in the Methods section (2) of the paper A comparison of background correction methods for two-colour microarrays. Kooperberg: Kooperberg et al. (2002) suggest an empirical Bayes model made up of a convolution of normal distributions to background adjust the signals. Observed foreground and background mean intensities (Xf and Xb) and their standard deviations (SDf and SDb), along with the number of foreground and background pixels (nf and nb) for each spot in a given channel are used in the model,